Purification and molecular characterization of recombinant rat betacellulin - Journal of Molecular Endocrinology

Page created by Rodney Bradley
 
CONTINUE READING
239

Purification and molecular characterization of recombinant
rat betacellulin
A J Dunbar, I K Priebe, M P Sanderson1 and C Goddard1
Cooperative Research Centre for Tissue Growth and Repair, CSIRO Health Sciences and Nutrition,
  PO Box 10065 Adelaide BC, South Australia 5000, Australia
1
 GroPep Limited, PO Box 10065, Adelaide BC, South Australia 5000, Australia
(Requests for offprints should be addressed to A J Dunbar; Email: andrew.dunbar@gropep.com.au)

ABSTRACT
A method for the large scale expression and                     Factor Xa cleaved an additional site within the BTC
purification of rat betacellulin (BTC) from                     protein, generating a truncated isoform separable
Escherichia coli has been developed using a cleavable           from full-length BTC by heparin-affinity chroma-
fusion protein strategy. Insoluble fusion protein               tography. Recombinant rat BTC stimulated the
collected as inclusion bodies was dissolved in urea             proliferation of mouse Balb/c 3T3 fibroblasts and
under reducing conditions, re-folded, and purified              competed for binding to the ErbB1 receptor in a
by gel filtration chromatography and C4 RP-HPLC.                dose-dependent manner analogous to that of BTC
Authentic rat BTC was obtained after proteolytic                purified from natural sources.
cleavage of the fusion protein with Factor Xa.                  Journal of Molecular Endocrinology (2001) 27, 239–247

INTRODUCTION                                                    initiation and transmission of highly conserved
                                                                signal transduction cascades (such as the mitogen-
Betacellulin (BTC) belongs to a class of growth                 activated protein kinase and phosphatidyl inositol
factors characterized by a six-cysteine consensus               3-kinase pathways), culminating in diverse cellular
motif that forms three intramolecular disulphide                effects, including growth, differentiation, migration
bonds critical for binding to the ErbB receptor                 and survival.
tyrosine kinase family (Dunbar & Goddard 2000).                    BTC was initially purified and characterized from
Collectively, this group of proteins are referred to as         the conditioned medium of a mouse pancreatic
the epidermal growth factor (EGF) family. Mam-                  -cell carcinoma cell line (Shing et al. 1993) and
malian members of this family identified to date                subsequently has been identified and characterized
include EGF, amphiregulin, and transforming                     in human (Sasada et al. 1993), bovine (Dunbar et al.
growth factor- (TGF-), which bind specifically to             1999) and rat (Tada et al. 2000). Increased
ErbB1, and heparin-binding EGF (HB-EGF),                        expression of BTC mRNA in the pancreas
epiregulin and BTC, which exhibit dual receptor                 compared with other tissues, and an increasing
specificity in that they bind both ErbB1 and ErbB4.             number of in vitro studies with cultured cell lines
The third group comprises the neuregulins (NRGs)                have supported the hypothesis that BTC signaling
which encompass the products of four genes                      through ErbB receptors plays an important part in
(NRG1–NRG4). Members of this sub-family bind                    islet growth and development in the pancreas. For
ErbB4 but not ErbB1, and can bind ErbB3 in the                  example BTC, together with activin-A, can convert
context of an ErbB2–ErbB3 heterodimer (Harari &                 populations of AR42J rat pancreatic tumour cells
Yarden 2000). Recently, a new member of the EGF                 into insulin-secreting cells (Mashima et al. 1996);
family has been described (Strachan et al. 2001).               BTC is required for the induction of insulin and
This molecule, termed epigen, has been shown to                 glucokinase gene expression in PDX-1-expressing
activate ErbB1. Characterization of the full ErbB               glucagonoma cells (Watada et al. 1996) and BTC is
receptor specificity of epigen, however, awaits                 able to mediate the proliferation and differentiation
further investigation. Binding and activation of the            of the rat insulinoma cell line, INS-1 (Huotari et al.
ErbB receptors by the EGF family results in the                 1998). Significantly, in these studies, neither EGF

Journal of Molecular Endocrinology (2001) 27, 239–247                        Online version via http://www.endocrinology.org
0952–5041/01/027–239  2001 Society for Endocrinology Printed in Great Britain
                                                                                      Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                                    via free access
240      and others ·       Recombinant rat betacellulin

      or TGF- could mimic these effects. In addition               University, College Station, TX, USA). Both cell
      BTC has been shown to exert a mitogenic effect on             lines were maintained in Dulbecco’s modified
      human undifferentiated pancreatic epithelial cells            Eagle’s medium (DMEM) supplemented with
      (Demeterco et al. 2000) and ductal epithelial cells          60 mg/ml penicillin, 100 mg/ml streptomycin,
      derived from developing pancreatic buds of rat               1 mg/ml fungizone and 10% (v/v) fetal bovine serum
      embryos (Sundaresan et al. 1998). Furthermore,               (FBS), and propagated at 37 C in a humidified
      immunohistochemical analysis has localized BTC to            atmosphere of 5% CO2.
      primitive duct cells in fetal pancreas and to some
      islet cell populations closely associated with insulin
                                                                   Cloning and expression
      producing cells (Miyagawa et al. 1999, Tada et al.
      1999). More recently, detailed analysis of ErbB1             The cDNA encoding the mature form of rat BTC
      deficient (-/-) mice has revealed that disruption of         (Asp32–Tyr111) was generated by RT-PCR. Total
      ErbB1 signaling leads to defects in pancreatic               RNA was isolated from 80–90% confluent rat IEC-6
      epithelial proliferation and an associated delay in          cells (ATCC CRL-1592) and cDNA synthesized
      -cell development (Miettinen et al. 2000).                  from 1 µg total RNA using oligo dT primer and
         In all cases described above, the evidence linking        Superscript II (Life Technologies, Melbourne,
      a role for BTC signaling to pancreas development             Australia). The subsequent cDNA was used as a
      and function through ErbB1 and possibly ErbB4 is             template for PCR with oligonucleotide primers, 5
      based almost solely on the use of cultured cell lines        ATC TAG GTT AAC ATC GAA GGT CGT
      in vitro. Clearly, analysis of animal models in which        GAT GGG AAC ACG ACC AGA ACC 3 (single
      the BTC gene has been either selectively eliminated          underline, HpaI restriction site; double underline,
      or overexpressed in the pancreas in vivo will be             nucleotide sequence encoding the Ile-Glu-Gly-Arg
      important, together with detailed investigation into         Factor Xa recognition site) and 5 CTA GAT AAG
      the effects of short- and long-term infusion of BTC           CTT TCA TCA GTA AAA CAG GTC CAC
      into animal disease models. Indeed, a recent study           CTG 3 (single underline, HindIII restriction site,
      by Yamamoto et al. (2000) revealed that admin-               double underline stop codons). The resultant
      istration of recombinant human BTC significantly             282 bp PCR product was purified, digested with
      improved glucose tolerance in mice with diabetes             HpaI/HindIII and cloned into HpaI/HindIII-
      induced by selective perfusion of alloxan. This              digested 46-amino acid porcine growth hormone
      effect was considered to be the result of promoting           (pGH(1–46)) expression vector (King et al. 1992) to
      -differentiation and regeneration from ductal or             generate pGH(1–46)-Ile-Glu-Gly-Arg–BTC (Fig.
      acinar cells, or both.                                       1). This vector was subsequently transformed into
         In this paper we report a method for the                  E. coli JM101. Large scale IPTG-induced protein
      large-scale production and purification of authentic         expression was performed in 41 litre fermenters
      recombinant rat BTC from Escherichia coli and the            (Applicon, Schiedam, The Netherlands).
      biological analysis of the purified protein. The
      ability to produce rat BTC in high yield and purity
                                                                   Fusion protein purification and re-folding
      will be particularly useful in examining the
      physiological effects of the administration of BTC            After fermentation, cells were disrupted by homog-
      to various rodent models of human disease                    enization, and insoluble inclusion bodies collected
      (including diabetes). Furthermore, these animal              by centrifugation (10 000 g, 25 min, 4 C). Inclusion
      studies can be performed without the complication            bodies were washed twice with 30 mM NaCl,
      of potential immunological ‘interference’ after              10 mM KH2PO4, 0·5 mM ZnCl2 and resuspended
      reaction to a heterologous protein (i.e. human               at 10% (w/v) in 100 mM Tris (pH 9·0), 8 M urea,
      BTC), particularly as we have recently demon-                40 mM glycine, 40 mM dithiothreitol (DTT) and
      strated that human BTC is highly immunogenic in              0·5 mM ZnCl2 and stirred for 30 min at room
      the rabbit (Bastian et al. 2000).                            temperature. Solubilized inclusion bodies were
                                                                   centrifuged (14 000 r.p.m., 20 min) to remove
                                                                   particulate matter and the supernatant was applied
      MATERIALS AND METHODS
                                                                   to a Cellufine GCL-1000 column (5100 cm)
                                                                   (Chisso Corp., Tokyo, Japan) equilibrated with
      Cell culture                                                 100 mM Tris (pH 9·0), 8 M urea, 40 mM glycine,
      Balb/c 3T3 and IEC-6 cells were from the American            40 mM DTT and 0·5 mM ZnCl2, at 2 ml/min.
      Type Tissue Culture Collection (Manassas, VA,                Fractions containing fusion protein (pGH46–rat
      USA) and AG2804 human lung fibroblasts were                  BTC) were pooled and subject to oxidative
      kindly donated by Dr J M Gunn (Texas A and M                 re-folding by diluting the mixture to a final protein
      Journal of Molecular Endocrinology (2001) 27, 239–247                                                 www.endocrinology.org
                                                                                    Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                                  via free access
Recombinant rat betacellulin ·    and others                     241

concentration of 0·1 mg/ml in buffer containing        ionization mass spectroscopy (Perkin-Elmer API
100 mM Tris (pH 9·0), 4 M urea, 40 mM glycine,        300, Shelton, CT, USA) and N-terminal sequence
5 mM EDTA, 0·4 mM DTT and 1 mM                        analysis (Hewlett-Packard G1000A, Palo Alto, CA,
2-hydroxyethylsulphoxide. After being stirred for     USA).
3 h at room temperature, the re-fold reaction was
terminated by pH adjustment to 4·0 with trifluor-
                                                      Mitogenic and ErbB1 receptor binding assay
acetic acid (TFA), filtered against a 1 µm membrane
to remove insoluble material and the mixture          Mitogenic activity of recombinant rat BTC (rrBTC)
applied to a C4 Prep-Pak RP-HPLC column               on Balb/c 3T3 fibroblasts was determined by
(40 mm100 mm; 300 Å, 15 µm; Millipore-               colourimetric assay using methylene blue as pre-
Waters, North Ryde, NSW, Australia) at 50 ml/         viously described (Dunbar et al. 1999). The ability
min. After extensive washing in 0·1% TFA the          of rrBTC to bind ErbB1 specifically was deter-
fusion protein was eluted with a gradient of 15–50%   mined by competitive displacement of 125I-
acetonitrile over 70 min (in the presence of 0·08%    recombinant human EGF (rhEGF) from ErbB1
TFA) at a flow rate of 25 ml/min. Fractions were      receptors present on AG2804 human lung fibro-
collected and those containing pure pGH46–rat         blasts. Briefly, AG2804 cells were grown to 70–80%
BTC fusion protein (as determined by SDS-PAGE)        confluence in DMEM/10% FBS in 24-well plates.
were pooled and lyophilized.                          The cells were washed twice with binding buffer
                                                      (100 mM Hepes, (pH 7·6), 120 mM NaCl, 5 mM
                                                      KCl, 1·2 mM MgSO4, 8 mM glucose, 0·1% BSA)
Factor Xa cleavage and purification                   and then incubated with 125I-rhEGF (labeled with
pGH46–rat BTC fusion protein was dissolved in a       Na[125I] using chloramine-T to a specific activity of
minimal volume of 10 mM HCl and resuspended in        approximately 20 µCi/µg) and increasing concen-
50 mM Tris chloride (pH 8·0), 100 mM NaCl,            trations of unlabelled rrBTC (0–44 nM) in binding
1 mM CaCl2 to give a final protein concentration of   buffer at 4 C for 18 h. Cells were then washed three
0·05 mg/ml. One hundred units of Factor Xa            times in Hanks’s buffered salt solution and lysed
(Amersham-Pharmacia Biotech., Castle Hill, NSW,       with 1 ml 0·5 M NaOH/0·1% (v/v) Triton X-100 for
Australia) were then added per 20 mg fusion protein   30 min. Radioactivities of cell lysates were then
and the reaction carried out for 20 h at room         determined with a -counter (Wallac 1470). Non-
temperature before terminating by pH adjustment       specific binding was determined by the addition of
to 3·0 with 1·5% TFA. Authentic rat BTC was           excess unlabelled rhEGF (100 nM) and was typi-
separated from the pGH46 fusion partner by            cally about 5% of total binding. Experimental data
RP-HPLC. The cleavage reaction was diluted 1:4        for both the mitogenesis assay and ErbB1 receptor
(v/v) with 0·1% TFA and applied to a C4 RP-HPLC       binding assay were fitted to a logistic four-
column (2·510 cm, 300 Å, Millipore-Waters) at        parameter dose–response model with variable slope
20 ml/min. The column was then washed with 0·1%       (SigmaPlot v.4·0).
TFA and BTC eluted from the column with a linear
gradient of 0–50% acetonitrile (in the presence of
0·08% TFA) at 25 ml/min over 80 min. Fractions        ErbB-1 receptor tyrosine phosphorylation
containing rat BTC were pooled and lyophilized.       assay
To separate full-length recombinant rat BTC from      AG2804 cells were grown to confluence in 10 cm
a shorter ‘mis-cleaved’ contaminant (see Results),    dishes and subsequently incubated for 12 h in
the protein pool was resuspended in 50 mM Tris        serum-free medium. Cells were then stimulated
chloride (pH 7·5) and applied to a Progel-TSK         with 10 nM rrBTC for 10 min at room temperature,
Heparin-5PW column (7·5 mm 7·5 mm, Supelco,          washed twice in PBS and then suspended in lysis
Castle Hill, NSW, Australia) attached to the HPLC     buffer (0·5 ml) (50 mM Tris chloride pH 7·4,
at a flow rate of 0·5 ml/min. The column was          150 mM NaCl, 1% deoxycholate, 1% Triton X-100,
washed with 50 mM Tris chloride (pH 7·5) and          0·1% SDS, 5 mM sodium orthovanadate, 10 mM
then eluted with a three-step gradient of 50 mM       sodium fluoride, 1 mM EGTA and complete
Tris chloride (pH 7·5)/0·2 M NaCl for 10 min,         protease inhibitors (Roche Biochemicals, Castle
followed by 50 mM Tris chloride (pH 7·5)/0·6 M        Hill, NSW, Australia). Cell lysates were cleared by
NaCl for 15 min, followed by 50 mM Tris chloride      centrifugation (20 min, 15 000 g at 4 C) and ErbB1
(pH 7·5)/0·2 M NaCl for 10 min. Fractions contain-    immunoprecipitated by incubating the lysate with
ing authentic full-length rat BTC were pooled,        1 µg anti-ErbB1 (Upstate Biotech., New York, NY,
desalted and the purity of the final preparation      USA) and protein-G sepharose (Roche Biochemi-
analysed by microbore C4 RP-HPLC, electrospray        cals) for 2 h at 4 C. Immune complexes were
www.endocrinology.org                                         Journal of Molecular Endocrinology (2001) 27, 239–247
                                                                           Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                         via free access
242      and others ·       Recombinant rat betacellulin

                         1. Description of the pGH(1–46)–rat BTC expression construct. The
                        downward arrow indicates the Factor Xa cleavage point. Amino acid and nucleotide
                        sequences of the various parts of the fusion protein are shown. The boxed amino
                        acid sequence corresponds to the complete amino sequence of rat BTC
                        (Asp32–Tyr111). Note: amino acid numbering is with respect to the full 177 amino
                        acids of the precursor rat BTC sequence. *Stop codon.

      collected by centrifugation, washed three times in           puncture. Rat BTC antisera was affinity-purified
      lysis buffer and heated (3 min, 95 C) in SDS-                using a High Trap protein-G sepharose column
      PAGE sample buffer. Proteins were separated on                (Amersham-Pharmacia Biotech.), following the
      6% SDS-PAGE gels and transferred to nitro-                   manufacturer’s instructions.
      cellulose filters (Hybond C, Amersham Pharmacia
      Biotech.). Blots were probed with anti-
      phosphotyrosine monoclonal antibody (PY20, Santa             RESULTS
      Cruz Biotech., Santa Cruz, CA, USA) and then
      with HRP-conjugated sheep anti-mouse IgG                     Production and molecular characterization of
      (Silenus Laboratories, Boronia, Australia). HRP-             rat BTC
      labeled proteins were visualized using enhanced
      chemiluminescence (ECL) (Amersham Pharmacia                  The mature form of rat BTC (Asp32–Tyr111) was
      Biotech.). To confirm equal loading, blots were              expressed as a recombinant fusion protein contain-
      stripped and re-probed with the anti-ErbB1                   ing the first 46 amino-terminal amino acids of
      antibody and HRP-conjugated rabbit anti-sheep                porcine growth hormone (pGH) (King et al. 1992).
      IgG (Zymed Laboratories Inc., San Francisco, CA,             A proteolytic cleavage site (Ile-Glu-Gly-Arg) was
      USA).                                                        engineered downstream of the pGH fusion partner
                                                                   to allow for ‘release’ of rat BTC after digestion with
                                                                   Factor Xa (Fig. 1).
      Production of anti-rat BTC antibody                             Insoluble pGH46–rat BTC within inclusion
      Antibody production was approved by the Animal               bodies was dissolved in urea/DTT-containing
      Ethics Committee of the Women’s and Children’s               buffer and partially purified by cellufine gel
      Hospital, Adelaide, South Australia and the pro-             filtration chromatography to remove endogenous
      cedure followed the Australian Code of Practice for          proteases. The solubilised fusion protein was then
      the care and use of animals for scientific purposes.         subject to oxidative re-folding in the presence of
      Briefly, three female semi-lop rabbits (obtained             2-hydroxyethyldisulphide. Fusion protein contain-
      from the Institute of Medical and Vetinary                   ing the correctly folded six-cysteine consensus EGF
      Sciences, Gillies Plains, South Australia) were              motif was then separated from mis-folded isomers
      injected with 500 µg of a 14-residue peptide                 by C4 RP-HPLC. pGH46–rat BTC fusion protein
      (G50ENCTGTTPRQKSK63) conjugated to dip-                      from this two-step procedure was considered to be
      theria toxoid (Chiron Technologies, Melbourne,               essentially homogeneous as judged by analytical
      Victoria, Australia) in Freund’s complete adjuvant,          HPLC and SDS-PAGE (Fig. 2).
      following standard procedures (Cooper & Petterson
      1999). After 6 weeks, animals were given a first
      booster injection of 100 µg peptide in Freund’s              Separation of the pGH fusion partner from
      incomplete adjuvant; they received a similar, second         authentic rat BTC
      boost, 4 weeks later. Blood samples were obtained 2          Incubation of the fusion protein with Factor Xa for
      weeks after each booster injection for determination         20 h at 22 C generated a complex profile of protein
      of antibody titre. Animals were killed 4 weeks after         peaks (P1–5) that were resolvable by C4 RP-HPLC
      the second booster and blood collected by cardiac            (Fig. 3A, B). P3 was identified as containing rat
      Journal of Molecular Endocrinology (2001) 27, 239–247                                                  www.endocrinology.org
                                                                                     Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                                   via free access
Recombinant rat betacellulin ·    and others                     243

 2. Analytical RP-HPLC analysis of purified
pGH46–rat BTC fusion protein. An aliquot of the
purified pGH46–rat BTC fusion protein preparation was
analysed on a microbore C4 RP-HPLC column (2·1 mm
100 mm, Brownlee Laboratories, Santa Clara, CA,
USA). Protein was eluted with an increasing gradient of
acetonitrile to 80% in the presence of 0·08% TFA over
37 min. Inset: SDS-PAGE analysis of the purified
preparation. Lane 1, molecular weight markers; lane 2,
pGH46–rat BTC fusion protein.

BTC by monitoring mitogenic activity (data not
shown). The identity and composition of the other
peaks were not investigated. The purity of the
recovered rat BTC (P3) was assessed by analytical
RP-HPLC (Fig. 3C), SDS-PAGE and electrospray
ionization mass spectrometry (Fig. 4A, B). A
protein of 9036·1±0·8 Da corresponding to full-
length rat BTC (theoretical Mr 9034·2 Da) was
identified. In addition however, a second protein
species of Mr 5992·00·4 was present in approxi-
mately equivalent amounts. N-Terminal sequence
analysis (five cycles) of the purified P3 preparation
identified two protein sequences: Asp-Gly-Asn-
Thr-Thr and Ser-Lys-Thr-His-Phe. Therefore,
Factor Xa, in addition to cleaving pGH46–rat BTC
fusion protein after Ile-Glu-Gly-Arg also cleaved
                                                           3. Analytical RP-HPLC of the pGH46–rat
an additional cryptic Factor Xa site within the BTC
                                                          BTC fusion protein cleavage reaction. Pure pGH46–
protein (27Pro-Arg-Gln-Lys?Ser-Lys-Thr33) (Fig.           rat BTC fusion protein was incubated for 20 h with
4B). The theoretical molecular mass of the ‘clipped’      Factor Xa at 22 C and an aliquot of the reaction
BTC isoform was predicted to be 5989·6 Da which           analysed by microbore C4 RP-HPLC as described in
is consistent with the observed molecular mass of         Fig. 2. (A) RP-HPLC analysis of the pGH46–rat BTC
5992·0 ± 0·4 Da. As expected, a polyclonal antibody       fusion protein at baseline (T=0 h). (B) Analysis of an
raised against a synthetic peptide corresponding to       aliquot of the reaction after 20 h (T=20 h). (C)
the amino acid sequence Gly-Glu-Asn-Cys-Thr-              Analysis of an aliquot of pooled fractions containing
Gly-Thr-Thr-Pro-Arg-Gln-Lys-Ser-Lys only rec-             P3 after large-scale resolution of the factor Xa
ognized the full-length BTC isoform and not the           peptides.
clipped form (Fig. 4A).
www.endocrinology.org                                             Journal of Molecular Endocrinology (2001) 27, 239–247
                                                                               Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                             via free access
244      and others ·       Recombinant rat betacellulin

                          4. (A) SDS-PAGE and western blot analysis of the purified P3
                         preparation. An aliquot of the P3 preparation was resolved by SDS-PAGE and
                         either stained with Coomassie blue (left panel) or immunoblotted and probed with
                         a polyclonal anti-rat BTC antibody to the peptide Glu-Asn-Cys-Thr-Gly-
                         Thr-Thr-Pro-Arg-Gln-Lys (right panel). (B) N-terminal sequence analysis (five
                         cycles) (double underline) and electrospray ionization mass spectrometry of the P3
                         preparation. Single underline indicates the amino acid sequence used to generate a
                         polyclonal anti-rat BTC antibody. ? indicates the cryptic Factor Xa cleavage site
                         within the BTC protein.

        Heparin-affinity       chromatography     effectively          phosphorylation of the ErbB1 receptor was analysed
      separated the full-length and clipped BTC isoforms            by western blotting using anti-phosphotyrosine
      (Fig. 5A, B). Full-length BTC eluted slightly earlier         antibodies. rrBTC clearly induced ErbB1 phos-
      than the clipped form (0·32 M and 0·4 M NaCl                  phorylation compared with control (Fig. 6A, inset).
      respectively). Fractions collected from the heparin-          Furthermore, binding of rrBTC to ErbB1 receptors
      affinity column containing full-length rat BTC were             also stimulated the proliferation of Balb/c 3T3 cells
      pooled, desalted and analysed by analytical RP-               in a dose-dependent fashion (Fig. 6B).
      HPLC, electrospray ionization mass spectrometry
      and N-terminal sequence analysis. The purified
      peptide eluted as a single symmetrical peak
                                                                    DISCUSSION
      (Fig. 5C) with a molecular weight of 9036 Da.
      N-Terminal sequence analysis confirmed the pres-
      ence of only a single protein sequence, correspond-           Human BTC has previously been produced in
      ing to rat BTC.                                               E. coli as a recombinant protein encompassing the
                                                                    mature peptide (Asp32–Tyr111). To initiate trans-
                                                                    lation of this protein, a methionine residue was
      Biological activity of rrBTC                                  appended to the amino-terminus (Asp32) (Seno
      rrBTC competed with 125I-rhEGF for binding to                 et al. 1996). Here we have sought to produce, in
      ErbB1 receptors present on AG2804 cells in a                  large yields, a rat BTC of which the authenticity
      dose-dependent fashion; 50% inhibition of binding             would not be compromised by the presence of
      (IC50) of 125I-rhEGF was observed at 0·68                    additional amino acids at the amino-terminus. To
      0·05 nM (mean..) (Fig. 6A). To determine                   do this, the mature form of rat BTC (Asp32–Tyr111)
      whether binding to the extracellular domain of                was expressed in E. coli as a fusion protein in which
      ErbB1 activated the receptor, AG2804 cells were               a site-specific proteolytic cleavage site was en-
      incubated with or without rrBTC. After immuno-                gineered downstream of the fusion partner. This
      precipitation with an anti-ErbB1 antibody, tyrosine           allowed for the generation of authentic BTC
      Journal of Molecular Endocrinology (2001) 27, 239–247                                                    www.endocrinology.org
                                                                                       Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                                     via free access
Recombinant rat betacellulin ·    and others                       245

                                                          6. Biological activity of rrBTC. (A) Binding of
                                                         rrBTC to ErbB1 receptors present on AG2804 cells.
 5. Heparin-affinity chromatography. The P3          125
                                                            I-rhEGF was added to AG2804 cells with increasing
preparation was further resolved on a Progel-TSK
                                                         amounts of rrBTC. Inset: ErbB1 receptor phosphoryl-
Heparin-5PW HPLC affinity column (A). Fractions
                                                         ation. AG2804 cells were treated with or without rrBTC
were analysed by SDS-PAGE (B) and those containing
                                                         and ErbB1 receptor phosphorylation examined by
pure rat BTC were pooled. Pure rat BTC ran as a single
                                                         immunoprecipitation and western blotting with anti-
symmetrical peak when analysed by microbore C4
                                                         phosphotyrosine (p-Tyr) antibody. (B) Promotion of
RP-HPLC with a gradient of 0–50% acetonitrile over
                                                         Balb/c 3T3 cell proliferation in the presence of rrBTC:
20 min (C).
                                                         cell proliferation is expressed as the % increase in absorb-
                                                         ance (655 nm) above control lacking growth factor. Data
                                                         in (A) and (B) are means ± .. for triplicate determina-
without the need to insert an amino-terminal             tions and are fitted to a logistic four-parameter dose–
methionine residue for translation initiation.           response model with variable slope (SigmaPlot v.4·0).
  Like many other proteins that are overproduced
in E. coli, pGH46–rat BTC fusion protein aggre-
gated in the cell to form insoluble inclusion bodies.    centrifugation in which the inclusion bodies
This insolubility made it possible to partially purify   segregated to the pellet phase. From 4 litres of
the protein from lysed cells by a simple low-speed       induced culture, we typically obtained about
www.endocrinology.org                                             Journal of Molecular Endocrinology (2001) 27, 239–247
                                                                                Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                              via free access
246      and others ·       Recombinant rat betacellulin

      35 g (wet weight) of inclusion bodies, contain-              Biological activity of rrBTC
      ing approximately 1 g of fusion protein. After               The biological activity of the rrBTC preparation
      dissolution of inclusion bodies and gel filtration to        was evaluated by both ErbB1 receptor binding and
      remove endogenous proteases, pGH46–rat BTC                   phosphorylation, and mitogenic stimulation of
      fusion protein was re-folded in the presence of              mouse fibroblasts. Recombinant rat BTC competed
      2-hydroxyethyldisulphide, and correctly folded               with [125I]-rhEGF for binding to ErbB1 receptors
      protein containing the six-cysteine consensus EGF            present on AG2804 cells in a dose-dependent
      motif (C1-C3, C2-C4 and C5-C6) separated from                fashion; 50% inhibition of binding (IC50) of
      mis-folded 1-, 2- and 3-disulphide scrambled                 125
                                                                      I-rhEGF was observed at 0·68 ± 0·05 nM,
      isomers by C4 RP-HPLC.                                       consistent with the findings of previous studies
         The blood coagulation Factor Xa was then used to          examining human BTC binding to CHO cells
      separate the pGH fusion partner from authentic rat
                                                                   expressing ErbB1 (IC50 =1 nM; Pinkas-Kramarski
      BTC. In addition to cleaving the fusion protein at           et al. 1998). Binding of rrBTC to ErbB1 also
      the canoconical tetrapeptide sequence Ile-Glu-Gly-           induced receptor phosphorylation in AG2804 cells
      Arg, Factor Xa also efficiently cleaved a non-                 and stimulated proliferation of Balb/c 3T3 cells in a
      specific peptide sequence (27Pro-Arg-Gln-Lys?                dose-dependent fashion (Fig. 6B). Interestingly, the
      Ser-Lys-Thr33) within the N-terminal region of rat
                                                                   clipped BTC also stimulated the proliferation of
      BTC. The non-specific Factor Xa cleavage event did
                                                                   Balb/c 3T3 cells with a similar dose-dependency
      not appear to be a secondary consequence of ex-
                                                                   (data not shown), confirming previous reports
      tended incubation, as a time course analysis demon-
                                                                   that the amino-terminal 30 amino acids are
      strated that the generation of both peptides occurred
                                                                   dispensable for ErbB1 receptor binding and
      simultaneously (data not shown). Interestingly,
                                                                   activation (Watanabe et al. 1994).
      recombinant human BTC produced in mouse A9
                                                                      The expression system that we describe provides
      cells is sensitive to endogenous proteolytic cleavage
                                                                   an efficient and consistent means of producing
      in an identical position (27Thr-Gln-Ser-Lys?Arg-
                                                                   highly purified and biologically active full-length
      Lys-Gly33), suggesting that this region within both
                                                                   recombinant rat BTC. Authentic rat BTC will be
      human and rat BTC is overly susceptible to proteo-
                                                                   useful in examining the effect(s) of the admin-
      lytic cleavage (Watanabe et al. 1994).
                                                                   istration of this growth factor on various tissues
         Factor Xa is primarily specific for the Ile-Glu-
                                                                   (such as the pancreas) in normal rodents and various
      Gly-Arg recognition sequence, but it is also known
                                                                   rodent models of human disease.
      to cleave additional sites. Only a few of these
      additional sites have been characterized, and they
      have little in common with the Ile-Glu-Gly-Arg               ACKNOWLEDGEMENTS
      sequence: for example, Cys-Asn-Gly-Arg?Trp-Val
      (Nambiar et al. 1987), Ser-Leu-Ser-Arg?Met-Thr               We thank Sam Randles for technical assistance,
      (Quinlan et al. 1989), and Ala-Leu-Ala-Arg?Lys-              Jelle Lahnstein of the SRC for Basic and Applied
      Tyr (Nagai & Thøgresen 1987). In some cases,                 Plant Molecular Biology for N-terminal sequence
      where cleavage occurs immediately after lysine               analysis and Yoji Hayasaka (The Australian Wine
      residues, as is the case here, reversible acylation by       Research Institute) for electrospray ionization
      incubation with 3,4,5,6-tetrahydrophthalic anhy-             mass spectrometry. The Australian Federal Gov-
      dride is effective in blocking non-specific cleavage          ernment Cooperative Research Centres Programme
      (Wearne 1990).                                               supported this work.
         To separate full-length rat BTC from the clipped
      shorter form, we initially used RP-HPLC. How-
      ever, even using very shallow gradients we were              REFERENCES
      unable to achieve suitable separation of the two
      forms (data not shown). In contrast, heparin-affinity          Bastian SEP, Dunbar AJ, Priebe IK, Owens PC & Goddard C
      chromatography effectively separated the two forms              2000 Measurement of betacellulin levels in bovine serum,
      (Fig. 5A, B). Full-length rat BTC eluted as a single           colostrums and milk. Journal of Endocrinology 168 203–212.
                                                                   Cooper HM & Petterson Y 1999 Production of polyclonal
      peak from an analytical C4 RP-HPLC and was                     antisera. In Current Protocols in Molecular Biology, vol I,
      shown to be greater than 99% pure as judged by                 pp 16·2·1–16·2·5. Eds JS Bonifacino, M Dasso, JB Harford,
      electrospray ionization mass spectrometry and                  J Lippincott-Schwatrz & KM Yamada. New York:
      N-terminal sequence analysis. Using this procedure             John Wiley.
                                                                   Demeterco C, Beattie GM, Dib SA, Lopez AD & Hayek A
      we were typically able to generate approximately               2000 A role for activin A and betacellulin in human fetal
      10–15 mg pure rat BTC from 4 litres of starting                pancreatic cell differentiation and growth. Journal of Clinical
      culture.                                                       Endocrinology and Metabolism 85 3892–3987.

      Journal of Molecular Endocrinology (2001) 27, 239–247                                                     www.endocrinology.org
                                                                                        Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                                      via free access
Recombinant rat betacellulin ·    and others                        247

Dunbar AJ & Goddard C 2000 Structure–function and                     Sasada R, Ono Y, Taniyama Y, Shing Y, Folkman J &
  biological role of betacellulin. International Journal of             lgarashi K 1993 Cloning and expression of cDNA encoding
  Biochemistry and Molecular Biology 32 805–815.                        human betacellulin, a new member of the EGF family.
Dunbar AJ, Priebe I K, Belford DA & Goddard C 1999                      Biochemical and Biophysical Research Communications 190
  Identification of betacellulin as a major peptide growth factor       1173–1179.
  in milk: purification, characterization and molecular cloning       Seno M, Tada H, Kosaka M, Sasada R, Igarashi K, Shing Y,
  of bovine betacellulin. Biochemical Journal 344 713–721.              Folkman J, Ueda M & Yamada H 1996 Human betacellulin,
Harari D & Yarden Y 2000 Molecular mechanisms underlying                a member of the EGF family dominantly expressed in
  ErbB2/HER2 action in breast cancer. Oncogene 19                       pancreas and small intestine, is fully active in a monomeric
  6102–6114.                                                            form. Growth Factors 13 181–191.
Huotari M-A, Palgi J & Otonkoski T 1998 Growth factor-                Shing Y, Christafori G, Hanahan D, Ono Y, Sasada R,
  mediated proliferation and differentiation of insulin-                 Igarashi K, Folkman J 1993 Betacellulin: a mitogen from
  producing INS-1 and RINm5F cells: identification of betacel-          pancreatic beta cell tumors. Science 259 1604–1607.
  lulin as a novel -cell mitogen. Endocrinology 139 1494–1499.       Strachan L, Murison JG, Prestidge RL, Sleeman MA, Watson
King R, Wells JRE, Krieg P, Snoswell M, Brazier J, Bagley               JD & Kumble KD 2001 Cloning and biological activity of
  CJ, Wallace JC, Ballard FJ, Ross M & Francis GL 1992                  epigen, a novel member of the EGF superfamily. Journal of
  Production and characterization of recombinant insulin-like           Biological Chemistry 276 18265–18271.
  growth factor-1 (IGF-1) and potent analogues of IGF-1,              Sundaresan S, Roberts PE, King KL, Sliwkowski MX &
  with Gly or Arg substituted for Glu3, following their                 Mather JP 1998 Biological response to ErbB ligands in
  expression in Escherichia coli as fusion proteins. Journal of         nontransformed cell lines correlates with a specific pattern of
  Molecular Endocrinology 8 29–41.                                      receptor expression. Endocrinology 139 4756–4764.
Mashima H, Ohnishi H, Wakabayashi K, Mine T, Miyagawa J,              Tada H, Sasada R, Kawaguchi Y, Kojima I, Gullick WJ,
  Hanafusa T, Seno M, Yamada H & Kojima I 1996                          Salomon DS, Igarashi K, Seno M & Yamada H 1999
  Betacellulin and activin A coordinately convert amylase-              Processing and juxtacrine activity of membrane-
  secreting pancreatic AR42J cells into insulin-secreting cells.        anchored betacellulin. Journal of Cellular Biochemistry 72
  Journal of Clinical Investigation 97 1647–1654.                       423–434.
Miettinen PJ, Huotari M, Koivisto T, Ustinov J, Palgi J,              Tada H, Seno M, Yamada H, Sasada R & Igarashi K 2000
  Rasilainen S, Lehtonen E, Keski-Oja J & Otonkoski T 2000              Molecular cloning and expression of rat betacellulin.
  Impaired migration and delayed differentiation of pancreatic           Biochimica et Biophysica Acta 1492 285–288.
  islet cells in mice lacking EGF-receptors. Development 127          Watada H, Kajimoto Y, Myagawa J, Hanafusa T, Hamaguchi
  2617–2627.                                                            K, Matsuoka TA, Yamamoto K, Matsuzawa Y, Kawamori R
Miyagawa J-I, Hanafusa T, Sasada R, Yamamoto K, Igarashi                & Yamasaki Y 1996 Pdx-1 induces insulin and glucokinase
  K, Yamamori K, Seno M, Tada H, Nammo T, Li M,                         expressions in -TC 1 clone 6 cells in the presence of
  Yamagata K, Nakajima H, Namba M, Kuwajima M &                         betacellulin. Diabetes 45 1826–1831.
  Matzuzawa Y 1999 Immunohistochemical localization of                Watanabe T, Shintani A, Nakata M, Shing Y, Folkman J,
  betacellulin, a new member of the EGF family, in normal               lgarashi K & Sasada R 1994 Recombinant human
  human pancreas and islet tumor cells. Endocrine Journal 46            betacellulin. Molecular structure, biological activities, and
  755–764.                                                              receptor interaction. Journal of Biological Chemistry 269
Nagai K & Thøgresen HC 1987 Synthesis and sequence-                     9966–9973.
  specific proteolysis of hybrid proteins produced in                 Wearne SJ 1990 Factor Xa cleavage of fusion proteins.
  Escherichia coli. Methods in Enzymology 153 461–481.                  Elimination of non-specific cleavage by reversible acylation.
Nambiar KP, Stackhouse J, Presnell SR & Benner SA 1987                  FEBS Letters 263 23026.
  Expression of bovine pancreatic ribonuclease A in Escherichia       Yamamoto K, Miyagawa J-I, Waguri M, Sasada R, Igarashi K,
  coli. European Journal of Biochemistry 16 67–71.                      Li M, Nammo T, Moriwaki M, Imagawa A, Yamagata K,
Pinkas-Kramarski R, Lenerink AEG, Bacus SS, Lyass L, van                Nakajima H, Namba M, Tochino Y, Hanafusa T &
  de Poll MLM, Klapper LN, Tzahar E, Sela M, van Zoelen                 Matsuzawa Y 2000 Recombinant human betacellulin
  EJJ & Yarden Y 1998 The oncogenic ErbB-2/ErbB-3                       promotes the neogenesis of -cells and ameliorates glucose
  heterodimer is a surrogate receptor of the epidermal growth           intolerance in mice with diabetes induced by selective alloxan
  factor and betacellulin. Oncogene 16 1249–1258.                       perfusion. Diabetes 49 2021–2027.
Quinlan RA, Moir RD & Stewart M 1989 Expression in
  Escherichia coli of fragments of glial fibrillary acidic protein:
  characterization, assembly properties and paracrystal                20 March 2001
  formation. Journal of Cell Science 93 71–83.                         21 June 2001

www.endocrinology.org                                                           Journal of Molecular Endocrinology (2001) 27, 239–247
                                                                                              Downloaded from Bioscientifica.com at 11/27/2020 01:34:41AM
                                                                                                                                            via free access
You can also read